Associations of cumulative exposure and dynamic trajectories of the combined Chinese visceral adiposity index and frailty index with incident cardiometabolic multimorbidity: a longitudinal analysis based on CHARLS

Abstract Background Visceral adiposity and frailty may jointly identify metabolic and physiologic vulnerability, but their longitudinal relationship with cardiometabolic multimorbidity (CMM) is uncertain. We examined initial, cumulative, and two-time-point profile measures of the Chinese visceral adiposity index (CVAI), a 32-deficit frailty index (FI), and their product in relation to incident CMM. Methods This landmark cohort study used China Health and Retirement Longitudinal Study data. CVAI and FI were measured in 2011 and 2015; participants with CMM by the 2015 landmark were excluded. CMM was defined as the coexistence of at least two of diabetes, heart disease, and stroke. Initial values, time-weighted cumulative exposure, and K-means two-time-point profiles were evaluated. Cox models estimated hazard ratios (HRs) after adjustment for sociodemographic, behavioral, anthropometric, kidney, lipid-treatment, hypertension, and medication variables. Restricted cubic splines, subgroup analyses, competing-risk and early-event exclusions, additional metabolic adjustment, and multiple-imputation analyses assessed robustness. Conventional receiver operating characteristic (ROC) and decision-curve analyses were exploratory descriptive assessments. Results Among 5,150 adults free of CMM at the 2015 landmark, 491 developed CMM over a median 5.08 years. In fully adjusted complete-case models, each 1-SD increment in initial CVAI, FI, and the CVAI-FI product was associated with HRs of 1.34 (95% CI, 1.15–1.56), 1.32 (1.21–1.45), and 1.27 (1.18–1.38), respectively. Corresponding HRs for cumulative measures were 1.43 (1.19–1.73), 1.47 (1.34–1.60), and 1.42 (1.30–1.54). Compared with the lowest product-profile class, the intermediate and highest classes had HRs of 2.74 (2.15–3.49) and 2.96 (2.07–4.23). Participants jointly in the highest CVAI and FI tertiles had an HR of 8.26 (3.49–19.53) versus those jointly in the lowest tertiles. Product associations were nonlinear. The cumulative product had an exploratory ordinary ROC area under the curve of 0.71 (0.69–0.74). Results were materially unchanged across sensitivity analyses and five imputed data sets. Conclusions Higher CVAI, frailty burden, and their co-occurrence before a 2015 landmark were associated with incident CMM. Repeated joint assessment may help characterize risk heterogeneity, but external validation and prediction-focused evaluation are required before clinical use.